Designing Handheld Devices That Don’t Hurt

(This article was first published under the title “What Separates Us” in the Spring 2015 issue of Innovation, the quarterly of the Industrial Designers Society of America. )

Other species can’t hold a toothbrush, turn a doorknob, grasp a hammer, hitch a ride, reach an octave or hit a space bar like we can.

Our ability to form a good grip using our thumbs and opposing fingers differentiates us, and as our tools have evolved from hammers to smartphones, we increasingly wield our evolutionary advantage by swiping, scrolling and texting. (When’s the last time you saw an orangutan play a video game?)

Over the past 20 years, the thumb has become an even more valuable player in our everyday lives. The dramatic increase in cell-enabled, GPS-enabled and other compact electronic devices has resulted in more handheld products that are designed to be thumb activated. Not only do these cameras, cellphones, game controllers and keypads cater to our thumbs, but they appear to be changing the way we use them. British writer Sadie Plant first mentioned this phenomenon in “On the Mobile,” her 2002 ethnographic research study on cellphone use. In it she quotes a participant who observes that teenagers even “point at things and ring doorbells with their thumbs.”

Obviously We’re Not Getting it Right

Unfortunately, the thumb is not wearing its popularity well. As the frequency of cellphone texting rises, the muscles in people’s thumbs get larger and subjective reports of pain associated with the activity increases. Ten years ago carpal tunnel syndrome, a wrist ailment, was common due to overuse of computer keyboards and mice. Today the diagnoses du jour are thumb conditions such as “Nintendinitis,” coined in Dr. Richard Brasington’s 1990 letter to the editor in the New England Journal of Medicine; texting thumb; and gamer’s thumb. Excessive texting has also been linked to other pathologies of the thumb, specifically joint arthritis and tendonitis. Obviously we’re not getting it right. What can we, as designers, do to provide safer products?

As product designers, we frequently hit a wall when designing for the hand. Anthropometric data is typically all we have at our disposal. Unfortunately, it does not help answer simple, vital questions, such as what can you reach with your thumb or fingers, or what can you grip comfortably? Here is where anthropometric data of the hand, and more specifically the thumb, is letting us down:

  • For the past 75 years, anthropometric data has come largely from military populations with much less data on civilians. The most comprehensive data relating to hand and finger anthropometry was based on US Air Force Flight personnel.
  • Anthropometric measurements in general are limited to individual body segments in isolation. This measurement technique does not capture the coordinated, multisegment/multijoint nature of more functional dimensions. As with other body parts, this is true of the hand where individual fingers (digits) are measured, but little functional data related to hand grip and finger reach is available.
anthropometrical data is often not useful for handheld device design

Imagine trying to design a handheld device with intended thumb actuation using this information, which is typical of the type found in anthropometry books.

  • The postures in which anthropometric dimensions are measured are typically unnatural and therefore not useful to designers. Joints are typically fully extended during measurement, not in natural, functional postures.
  • The collection and presentation of anthropometric data has focused on static poses and does not capture the dynamics of movement.
  • Anthropometric data, especially for the hand, is still most often presented in two-dimensional plots, numeric tables, calculators or templates. Three-dimensional range of motion is typically not well represented.
Humanscale 1/2/3 and handheld devices

Humanscale 1/2/3’s two-dimensional plots characterize the available anthropometric data, especially for the hand. Three-dimensional range of motion is not yet well represented.

Anthropometry’s generic offerings are less than useful when addressing specific design challenges with unique constraints. This tension between basic and applied sciences is common in other disciplines as well. In order for data to generalize to the largest possible audience (and be publishable), it has to feature the most nonspecific, unconstrained data. Hence the “basic” in “basic science.” Data collected for application to a specific set of design constraints is often seen as useless to anyone outside the immediate design team.

It’s possible anthropometry may catch up to our expanding need for data around hand grip and thumb reach. In the meantime, in order for us to ensure a good fit between thumb-actuated devices and their human users, we need different tools and methods.

Computer Modeling to the Rescue — Almost

In its 2003 book Kodak’s Ergonomic Design for People at Work, the company pioneered the notion that reach envelopes should be represented in three-dimensional space. Since then, computer models of three-dimensional anthropometric data have become available. These models have come a long way in making anthropometric data more useful and usable by allowing designers to envision the human body in relationship to product concepts while in the conceptual design phase — prior to building physical prototypes and while the design is still extremely malleable in CAD modeling software. That said, many anthropometric models still fail to adequately address the intricacies of the human hand and functional grip — specifically concerning thumb reach envelopes.

We need something to hold us over while we hope and wait for better anthropometric models of the human hand that will integrate well with our CAD systems. Reviewing the existing anthropometric data is always the best first step, but once you are convinced that the data you need is not available, I recommend moving on to collect your own product-specific data. This does not have to be super expensive or time consuming. While large sample sizes are always nice in anthropometric studies, a question that is tightly constrained, such as thumb reach on a specific product with a specific grip style by a specific user population, may be answered with data from as few as 30 to 50 people who fit the user profile.

Get creative when planning your study. Keep in mind that nobody has tackled the exact same design challenge and most likely nobody has collected the same data you are after.

A Real-World Example

A few years ago while working at Kodak, we designed a pocket video camera with a candy bar form factor. The marketing team defined two target user groups: teenage girls and mothers of young children. Our industrial designers asked a seemingly simple question about where to place buttons intended to be actuated by the user’s thumb on the back of the camera. We scoured the anthropometric literature and investigated the available computer models of the hand. None seemed adequate, so we decided to collect our own data. We had another study planned at that time on a related topic with the same two user groups. So we tacked the collection of some simple anthropometric data onto it.

Rather than simply measuring the length, breadth and depth of the thumb, we needed data that was directly applicable to our question about where to place controls. We needed to know what teenage girls and moms could reach on the surface of the camera. In order to describe this thumb reach envelope, we used an iPod Touch with a simple drawing program. To replicate people’s grips — a big impact on thumb reach — we built a small cradle for the iPod that extended its shape to the planned dimensions of the camera.

gathering thumb-reach data for handheld device design

Left, a thumb “painting” the iPod screen. Right, reach envelope data for the female teens and young mothers in our study.

We instructed participants to hold the iPod as if they were going to use it to capture a video. Then we asked them to swipe their thumbs across its surface. The drawing program running on the iPod captured their individual thumb reach envelopes as they “painted” the parts of the screen they could reach. Eventually, we combined the data from all participants in a group to determine the thumb reach envelope for that group — for example, specifying which parts of the screen 90% of participants could reach with their thumbs. We published this 2013 study, “Defining Thumb Reach Envelopes for Handheld Devices,” in Human Factors — not so much to suggest reuse of the data, but more to suggest that this method may be useful to others.

Bridge to the Future

While it is always fun, and often useful, to collect our own data, I hope the need for do-it-yourself thumb reach data does not last long. Surely, the companies working on three-dimensional models of thumb reach will grab others’ data and integrate it into their models. There has already been some progress toward using three-dimensional scanning of human hands, and researchers at Hokkaido University in Japan have proposed a system for virtual ergonomic assessment of products by integrating a digital hand model with a product model. The resulting anthropometric data will no doubt be merged with in-depth motion analysis of various styles of grip and typical hand motions.

All of this should culminate, eventually, in useful models of the hand that can be brought into standard desktop CAD applications to help designers determine how best to design products to avoid injury. And we will fondly recall this blip of time when the lack of readily accessible relevant data stuck out like a sore thumb.

 Checklist: Best Practices for Handheld Device Design 

  • Think beyond Humanscale 1/2/3. A table of numbers or a template cannot do justice to the complex postures and movements of the multiple articulated joints of the thumb and fingers. See if current computer models provide the anthropometric data you need in a format compatible with your CAD system.
  • Review existing anthropometric data, but once you are convinced that the data you need is not available, move on and collect your own data.
  • Get creative when planning your study. Keep in mind that nobody has tackled the exact same design challenge you face and most likely nobody has faced the challenge of collecting the data you need.
  • Consider sharing your results and techniques with others in a publication or conference talk.
  • Keep the pressure on your CAD system vendors for better plug-in modules for hand anthropometric data.